目 SciencVeoslu minhet 6Cp,:/Ioswludwe wa1.,sncFeadbr, aAuca.crDyi ,d2 0R14e gions 一 C ̄cation:Wang YJ,Lu PJ,Ma YZ,et a1.,2014.Response to climate change of different tree species and NDVI variation since 1923 in the middle arid region of Ningxia,China.Sciences in Cold and Arid Regions,6(1):30-36.DOI:10.372z!/SRJ.1226.2014.00030. ’Response to climate chan ̄e ol 。——to:limat,chan ̄,of lifferent tree pecies anddifferent tree specles and NDVI variation since 1 923 in the middle arid region of Ningxia,China YaJun Wang , ,RuiJie Lu 3,4YuZhen Ma 3,4HongWei Meng 3,4 ShangYu Gao , 7 ,CollegeofLifeSciences,BeijingNormalUniversity,Beijing100875,China 2.HebeiUniversiyotfEngineering,Handan,Hebei056038,China 3.KeyLaboratory ofEnvironmental Change andNaturalDisaster,Min&try ofEducation ofChina,BeijingNormal Universiyt .Beijing 100875,China 4.StateKeyLaboratoryofEarthSurfaceProcessesandResourceEcology,BeijingNormalUniversity,Beijing100875,China Correspondence to:YaJtm Wang,College of Life Sciences,Beijing Normal University.No.19,Xinjiekouwai Street,Beijing 100875,China.Tel:+86—10-58805730;E—mail:hbyjwang@163.corn Received:March 13,2013 Accepted:July 1,2013 ABSTRACT The normalized diference vegetation indcx eNDVT)is used extensively to describe vegetation cover and ecological environ— ment change.The purpose of mis study was to contrast the response of different tree species growing in the same habitat to climate change and retrieve Dast NDⅥusing tree—ring width data from tree cores collected from the transitiona1 zone of尸inus tabulae&rmis and Picea crassifolia in the Luoshan Mountains in het middle arid region ofNingxia.Correlation analysis indi— cated that radial growth of tabulaeformis is more sensitive to precipitation and temperature change than that of尸crass的h'a. Natural factors such as water availability and heat at this elevation arc more suited to the growth ofP crasMfolia.and are more advantageous to its renewal and succession.P crassifolia is probably the better of the two species for protecting the forest ecosystem and conserving water in the Luoshan desertiifcation area.Ring width of crassfolia correlates signiifcantly with average NDVl for April-May _0.641,P<(】.O1),and both ofthem are influenced positively by precipitation in April-May. he reconstTructed NDⅥfor 1 923-2007 shows the relatively 1ow vegetation cover occurred in the 1 920s一1 930s.the 1 960s-1 970s.and the early 2 1 st century.The reconstructed NDVI better reflected he droughtt climate in the study area. Keywords:Picea crassifolia;Pinus tabulaeformis;tree.ring;drought 1 Introduction (Tucker et a1.,l 986;Fang et a1.,2004;Luo et a1.,2009). However,this data.recorded since the l 980s.is unable to provide data about earlier environmental evolution. Tree.ring data have been broadly used in he study of tpaleoclimate,allowing the reconstruction of precipitation The normalized diference vegetation index(NDVI), defined by near-infrared and red reflectance remote sens— ing images.reflects changes in land vegetation(Wang et aL,2005;Tangeta1.,2006).Itis consideredtobethemost effective ndex fior monitoring regional nd aglobal vege— tation coverage and ecological environmental change, wih witde applications in vegetation remote sensing (Treydte et a1.,2006),temperature(Gurskaya et a1., 2012)。the Palmer Drought Severity Index(Cook et a1., 2010),streamflow(Cook nd aJacoby,1983)and sea sur. face temperature(D’Arrigo et a1.,1993).Studies on nat. YaJun Wang et al,2014/Sciences in Cold and Arid Regions,6(1):0030—0036 31 ural disasters by tree—ring data,such as glacier fluctuation growing in the same habitat,and revealed past (Luckman,1 993),volcanic explosions fBriffa,l 998) earthquakes(Sheppard and White,1 995) and pest plagues(Swetnam and Lynch,1 993)have alJ achieved good results.Hi 一resolution information about past en. vironmental change can also be reconstructed from hi曲一resolution NDVI changes by ring width of P Cras— sifolia.This provides information for the study of envi. ronmental changes in this fragile area and regianal re. sponse to global change,and offers material that may be beneficial to research on the dynamics of forest commu— nities and vegetation renewa1. 2 Materials and methods tree.ring data. The Luoshan Mountains,surrounded by desertiifed 1and,lies in the middle arid region of he Nitngxia Hui Autonomous Region,bordering the Tengger Desert and he Mu Ust Desert.The ecological environment in this region is rfagile and the vegetation growth is sensitive to climatic change.The reaction of tree radial growth to he TLuoshan Mountains(Figure 11 rae 1ocated inland. far fram the ocean,and he tclimate is d The main peak is 2,624.5 m a.s.1.,with an average elevation of 1.065 m. he verTtical distribution of the vegetation is distnctliy zoned wih tncrieasing elevation,from desert rassgland to shrubland,broadleaf forest,conifergoroadleaf forest and conifer forest.The sampling site was in a semi—arid rea awiIh comparatively low precipitation. precipitation(Wang et a1.,20 1 0)and he tPalmer Drought Severity Index(Wang et a1.。2013)has been studied.but there are no reports off the reaction ofdifferent tree spe— cies growing in the sanqe habitat to climate change and the relation between NDⅥand rtee—irng width.The pre. sent study explored the correlation between NDVI and ring width of P tabulaeformis and Picea crass{fO.i 05.5。E Z 1 06.0。E 1 06.5。E 1 070。E .1 07.5。E 1 9 。。 、-a J1chio Z 0 tt-5. 妻 、ID- z n ing r f 』 I uoshln Mt - c 、 ,, 、 China Z D-- - j,、’,,、,、、~ 、 , , ' J、 、 , 、 、 Legend - -三一i¨_ : , , _||r Tonj ; M㈣0I ,’ 、、 ▲ Sampling site 口NDVI re ̄ion ’ {mng ● Meteorological station 0 , 0 西 Zhangliayuan , , . 。 一一一Pray nc1al btmadarics Figure1 Locationsofstudyareaand sampling site he tTree-ring cores were collected in the upper dis. he CoFECHA tquality control program(Holmes.1 983). As a result.some cores were excluded due to low corre— lation coeficifents compared to the main sequence, tribution limit of P mbulaeformis(37。l8 N.106。16 E: 2,400 m)in the仃ansition zone where CraSS f 矗rst appears.These trees were 15—20 m in height,and the anomalous values.or short Jength.Chronologies were canopy ofP tabulacformis nd aP crassf f was 0. 一0.8 Many plant species were present in the shrub Iayer and herb layer beneath the trees.The ground layer was well developed。with a small accumulation of pine needles. ten esthablished using he tARSTAN program(Cook nd aHolmes,1 986).Taking into account the background of rtee growth vends produced by genetic characteristics,a negative exponential function(Fritts.1 976)was used to iftthetree growthfiendand removethefiend.Finally,the One Or two cores were taken from each of 20 P tabu— laeformis and 20 P crassifo{ia flees on a partly shaded slope. standard chranalogiesforbothP crassifotia(LS11 and tabulacformisfLs2)were developedfFigure2 Table 11_ Tree—ring theory generally posits that higher values 0f MS CC、and EPS indicate that the tree growth is sensitive to the change of climate factors;the samples selected contain more popul ̄ion signals of仃ee growth. Although the core number of he tLS 1 chronology is iust Following common procedures for tree—irng analysis (Stokes and Smiley,1968),the cores were dried,fixed ad polnished,and their width measured by a LINTAB irng width measurement device with a precision of 0.0I n1m.Cross—dating nd ameasured results were tested using 32 YaJun Wang et a1 2014/Sciences in Cold and Arid Regions,6(1):0030—0036 22 cores,according to MS,CC,and other statistics the cores selected can better represent the growth charac— 一鲁pI≥ 盛 I。0.I downloaded from the Environmental and Ecological Science Data Center for Wlest China.National Natural teristics of P crassifolia in the sampling site,and some information on environment change can be drawn from the LS 1 chronology. Science Foundation of China(http://westdc.westgis. ac.cn).The spatial resolution of the NDVJ database was 8 km and the temporal resolution was 1 5 days. The NDVI data used in this study were selected from the 0.5。×0.5。region containing the coring site.The re— gional NDV1 was obtained by averaging the NDVl val— ues of all he tpixels in the selected region.The maximum ofthe two NDVI values in one month was selected as the monthly NDVI value 厶 2.3 Climate data 量 日 Data for the monthly average temperature and monthly total precipitation around the study area for 1 98 1 007 was retrieved from the Tongxin Meteorolog— ∽ ical Station(36。59 N 1 05。54 E:l,345 m).Annual pre. cipiatiton was 267.0 mm:from June to August,it was 1 56.7 mm,or 58.7%of the annual precipitation most of which fe11 in August.The average annual temperature Ycar was 9.5。C:the highest temperatures occurred in July, averaging 23.3。C. Figure 2 Ring width chronology and sample depth.The solid line represents the yearly value and the dotted line indicates the sample depth;the arrow indicates the year for which the coeficifent of subsample 2.4 Methods A correlation analysis was performed to analyze the relation between NDVI.climate factors.and tree—ring signal strength fCsss)was>0.80 2.2 NDV/data NOAA/AVHRR NDVI 1 5一day fmaximum)compo— width,and a regression analysis was carried out to de— velop the conversion equation between the tree—ring width and NDVI.The equation was cross—validated fWalpo1 and Myers,1 985;Liu and Shao,2003)since the measured data spanned only 26 years. site data for August of 1981 tO December of 2006 were Table 1 Sample site information and statistics or fhe tchronologies Notes:CN refers to core number;MS is mean sensitivity;SD is standard deviation;CI is common interval,CC is correlation for all cores EPS is expressed population signal;Csss is the coeficientf ofsub—sample sinalg s ̄ength. 3 Results beginning to enter the wilted period.Most of the vegeta tion stopped growing in Novembe ̄giving an NDVI 3.1 NDVIchangeinoneyear close to hatt in May.The December and January NDVl values showed no measurable difference.The NDVI values from April to November were selected for the fo1 1owing analysis. 3.2 Correlations between NDVIandprecipitation,and betweenNDVIandtemperature The annual change trend of NDVI(Figure 31 shows that the NDVI values in January,February,and March slightly decreased.and he tminimum was in March.The April NDVI f0.1043)was slightly hiherg than that in March f0.1022) implying that the vegetation was begin. ning to grow:in Mav it increased more rapidly,and it began to show an obvious increase from June to August, when it reached its maximum value.From September the The correlations were analyzed between NDVI and monthly precipittiaon,and between NDVI and tempera— ture,for the current month and for the previous two NDVI value gradually reduced as the vegetation was YaJun Wang et a1.,2014/Sciences in Cold and Arid Regions,6(1):0030—0036 33 months.The carrelations between NDVI and precipita. tion illustrated that NDVl correlated more significantly with precipitation in the last month and the month be— orethe1asttfhanwithprecipitationinthe currentmonth. Vegetation growth has a lag response to precipitation: rain signiicafntly affected vegetation coverage in the ollowing month,and also had an obvious ifnfluence on coverage in the succeeding month.The carrelation to November(Figure 4c)was statistically signiifcant.The positive correlation coefficient with May precipitation was relatively high.Further analysis showed that tree—ring width correlated positively and significantly with spring precipitation.Wide rings develop when spring precipiation itncreases.The ring width carrelated negatively with temperature in most months,approaching the 0.05 confidence level in July. The correlation between tree..ring width of R tabu.. fFigure 4a)was clearly positive in most months except April.when NDVI negatively carrelated with precipita. tion in February,March and April.For example signif- icant positive correlation was found between the May NDVI and precipitation in March <0.O1),April <0.o 1),and May <0.05); during the June NDVI and precipitation in April <0.0 1),May <0.O 1),and June <0.05):between the July NDVI and precipita. tion in May <0.01)and June <0.05):between the August NDVI and precipitation in July;and during both the September NDVI and October NDVI and precipita. tion in August <0.01). 0.24 0.21 0.1 8 凸 z 0 1 5 0.12 0 09 Jan. Mar May Ju1. Sop Nov Month Figure 3 Annual monthly NDVI Due to the arid climate in this region.higher temper. atures cause increased water consumption through evap— oration and transpiration.and therefore less available moisture for vegetation growth and a lower NDVI due to the reduced growth rate.The results of carrelations be wteen NDVI and temperature(Figure 4b1 showed that NDⅥhad a negative correlation with temperature in most months.reaching the 0.05 confidence 1evel in July and October.NDVI negatively correlated with tempera— ture jn either one or wto months earlier:for example,the NDVI in June correlated with temperature in May;the NDVI in August carrelated with temperature in July( <0.0 l 1:the NDVI in July correlated wiht temperature in May;and the NDVI in September correlated with tem— perature in Ju1y(p<0.05). 3.3 Correlations between tree—ring width andprecipi- tation,andbetween tree-ring width andtemperature Neither hte correlation between tree—ring widths ofr P crassifolia and precipitation nor temperature from April 1aeformis and precipitation,and between tree.ring width and temperature,from April to November(Figure 4d) showed that ring width correlated positively with precip— itation in most months except September,and reached the 0.05 confidence level in May.Tree—ring width negatively correlated with temperature in most months.and signifi. cantly correltaed with temperature in MaV <0.0 1)and July <0.01).It can be seen that more rain and lower temperatrues in the growth season in the stLldy area were advantageous to radial growth ofP atbulaeformis. In summary,radia1 growth of R tabulaeformis is more sensitive to precipitation and temperature change than p crass旅)t S,namely,the hydrothermal condi— tions in this zone better meets the growth needs of the latter.P crass )/ iS probably the better of the two spe— cies for protecting the forest ecosystem and conserving water in the I uoshan desertiifcation area. 3.4 Correlation between NDVIandtree-ring width Tree—irng widths of P crass帕妇had a signiifcantly positive correltaion wiht NDVI(Figure 4e)in May <0.0 l、:in particular,the correlation coe伍cient wiht average NDVI ofrApri】一Maywas maximal(r=0.641 P<0.001). The correlation between the April-November NDVI and tree—irng width for tabulaeformis was analyzed (Figure 4f).The positive correlations between ring witdh and NDVI in May,June,July,August,and October were at the 0.01 confidence Ieve1.Further analysis revealed that the correlation coemcient between ring width and average NDVI for April to August was 0.693 <0.00 1). Tree.ring width for P tabulaeformis signiifcantly correlated with NDVI during the peak growth period.For P crassi l tree.ring width correlated significantly with NDVI the early growth season.Tree.irng widths ofP atbu&eformis and NDVI for April-August all signiif— cantly correlated with precipitation for April-August, indicating that appreciable precipitation in the growth season is beneficial to vegetation growth.as might be expected. Tree—ring widths of P crassi l{n and April-May NDVI all positively correlated with precipitation for April—MaV-indicating that wide rings and high NDVI values correspond to greater spring rainfal1 and provide a reasonable explanation for the signiifcantly positive carrelation between tree.ring width and April-May NDV1. YaJun Wang et al,2014/Sciences in Cold and Arid Regions,6(1):0030-0036 Despite the fact that the correltiaon between ring width of尸crass , and precipitation did not reach the signiifcance leve1.correlation with July temperature was April-May NDVI fNDV ̄5).The conversion equation is NDVI4 =O.09+O.03X where X is tree.ring width.The correlation coeficient of fsigniifcant(r=一0.443,P<0.05) and he tcorrelation coef- ifcient between ring width nd aNDVI in April_lⅥaV was high(r=0.641,p<0.001).No reconstruction by irng width of P crassyblia and sfong uniformity in the trend of rtee—rng iwidth and April_IⅥav NDVI make us to extract NDⅥinformation from the change of ring width 0f P equation )is r=0.64 1 <0.OO 1);the explained vari— ance is 41.1%:the adiusted explained variance is 38.6%: and he ttest value F=16.07 <0.OO1) showing that the equation has a high significance.The reconstructed and the measuredNDVI are shown in figure 5. crassifolia.This also confnnsi the universality ofapply. ing tree—ring data to the reconstruction of past envkon- menta1 conditions. Cross—validation was applied to test the stability of the equation and reliability of the reconstruction using a number of test parameters.The reduction error fRE)is a 3.5 The establishment of a conversion equation be- tween tree-ring width andNDVI reliable statistic that precisely examines the climate re— construction and is a more effective diagnostic tool than other test sattistics(Wu.1 990).In this study RE=0-29. For the present study,the ree-ritng widths ofP crassi- greaterthan 0.which shows that equation f1) s a feasible means ofreconstructing NDVl4_ change. 加l were selected as the predictor for reconstuctring 琶 o 7 0 7 鼋 器 : 8 O 1 委一0.1 (c) +●■一 1 emperaturc P recipitatio ̄ ().5 ,(d) +Ptecipitatlon .一\一// \ .O.3 O.1 0.1 : .-J"[・cmperatur.c 芒: _-1- 。/- 0.5 0 3 0.3 0.5 0.7 ().7 0 6 0一o 7 --__ — 0 5 0 4 O3 O 2 O.1 0.O Figure 4 Correlation coeficfients between monthly NDVI and climate data and between monthly NDVI nd atre-ring width:(a)NDVI and precipitation;(b)NDVI and temperature;(c)LS 1 and precipitation,temperature;(d)LS2 nd aprecipitation,temperature; (e)LS1 andNDVI;(t)LS2 andNDVI. refers to 0.O1 confidence level; refers to 0.05 confidence level 3.6 Analysis froeconstructedND The year for the coefncient of he subsamplte signal > Q Z strength(Wigley et a1. 1984),CSSS>0.80 was selected as the beginning of the reconstruction.This sudy trecon— structed the NDVI4_ from 1 923 to 2007(Figure 6). Figure 6 shows that changes have obviously taken place since 1923 in the Luoshan Mountains.Points below the average NDVI represent a relatively low vegetation Year cover,and those above the average represent a relatively Figu1.e5 ConWastbetweenthereconsmactedandthemeasuredNDVI high vegetation cover.Thus,relatively high vegetation YaJun Wang eta1.,2014/Sciences in Cold andArid Regions,6(1):0030—0036 35 cover occurred in l93 1939. 1944-l965. and 1 979—1 999.and there was relatively low vegetation cover NDVI4_5 reconstruction since 1 923 was based on tree.ring widths from P cFassifol/a.It shows that there was relatively more vegetation in the study area in in I925—1935,1940-1943,1966—1978,and 2000__2005. The area experienced two signiifcant periods of ncreas.i ng vegetiation:1928-l 961 and 1975-1 986.and two of 1936-l939.1944-1965.and l979L一1999 with a humid climate,and less vegetation in 1925-l 935,l940-1943, decreasing vegetation:1 96 1 974 and 1 987-2003. () Figure 6 Reconstruction ofNDVb_s showing ifve-year moving averages.The thin line shows yearly value and the smooth line represents the ifve—year moving average.Horizontal line indicates overall average hTe relatively low vegetation coverage in the 1920s一1930s.the 1960s-1970s.and the early 2lst cen. tury was represented in the Palmer Drought Severity In— dex(PDSI1 changes for Mav—uly in the Helan Moun. tains duringthose periods(Chenet f 、2010)、also inthe low PDSI for May-June in the Changling Mountains (Chen et a1..2O1 l1.It can be seen that there is some syn. chronization between NDⅥvariation in the Luoshan Mountains and climatic modiifcation in the surrounding areas. hTe low NDVI in 1966-1 977 in the study area cor— responds to the high degree ofaridiyt ni the 1960s to the 1970s nad throughout northern China from l950 to 2000 (Wang and Zhai.2003).The low NDVI also coincides with the weakening of the Indian summer wind in the mid.1 9605 and late 1 970s(He gt aL.2005).The probable reason is that Indian Low Pressure abates when the Indi. an summer monsoon is weak and water vapor from low latitudes reduces,so the precipitation in the study area decreases.The comparison between NDVI reconstruction and the East Asina Summer Monsoon Index(EASMI) (http://web.1asg.ac.cn/staff/ljp/nidex.html1 shows that the low NDVI in the 1960s conforms to the low EASⅣⅡin the 1960s to the 1970s.The environment change in study area is stamped by global climate evolution. 4 Conelusions hTe NDVI of ground vegetation in the Luoshan Mountains is strictly limited by the amount of precipiat. tion,especially during the two months prior to growth, allowing ofr the lag time between rainfall and vegetation rgowth.Tlree-irng widths for P tabulacformis signiif. cantly correltaed with the NDVI for April-August =0.693,P<0.01),and P crassoColia correlated with the NDVI inApril-May( ,NDVI4-5) =0.641,P<0.01). 1 966-1 978.and 200 _2005 with a dry climate. Acknowledgments: Authors thank reviewers for their valuable comments. This research was supported by the National Natural Science Foundation of China(No.41 17l 159).and the Key Program ofthe Hebei Education Department(No. ZH20l2035). References Bri硒K&Jones PD.Schweingruber FH.et aL 1 998.Influence of volcanic emptions on Northem Hemisphere summer temperature over the past 600 years.Nature.393f41:450_1454.D0I: 10.1038/30943. Chen E Yuan YJ.Wlej WS, ,20l0.Long.term drought severity variations recorded in tree rings in the northem Helan Mountains. Advances in Climate Change Research 6r51:344—348. Chen F Yuan YJ,Wei WS, a1.,201 1.Reconsturction of May—June Palmer Drought Severiyt Index at South Margin of Tengger Desert, China since A.D.1691.Scientia Geographica Sinica.31f4): 434广1439. Cook ER Anchukaitis KJ,Buckley BM, aL,2010.Asian monsoon failure and megadrought during the 1ast millennium.Science, 328(5977):486__489.D0I:10.1l26/science.1185l88. Cook ER Holmes RL.1986.Users Manua1 for Program ARS_l’AN. Laboratory of丁ree—ring Research.Univ.ofArizona,Tucson.AZ. Cook ER.Jacoby GC.1983.Potomac River streamflow since 1730 as reconstructed by tree rings.Journal of Climate and Applied Mete— orology, 22(10): 1659—1672.http://dx.doi.org/10.1175/ 1520一O450r19831022<1659:PRSSAR>2.0.CO;2. D’ArrigoRD。CookER JacobyGC,etaL,1993.NAOand sea sruface temperatm-e signatures in tree—ring records from the North Atlantic sector.Quatemary Science Reviews.12:431—.44O.D0I:10.1007/ BF03 1 82768. Fang JY,Piao SL,He JS,et aL,2004.Increasing terrestrial vegetation activity in China,1982-1999.Science in China(Series C:Life Sci. ences).47f3):229—24O.D0I:10.1007/BF03182768. Fritts HC.1976._rree Rings and Climate.Academic Press,London. Gurskaya M,Hallinger M,Singha J,et a1.,2012.Temperatm'e recon- struction in hte Ob ver valley based on ring widths ofthree conif- erous tree species.Dendrochronologia,30:302-309. He L Wu BY Mao WX,2005.1'he interdecada1 v ab|ljty of nIdina summer monsoon and hte climate state shift in north China.Joumal ofTropica1 Meteorology,2 1 r3 :257-264. Holmes RL.1983.Computer-assisted quality control in rtee—ring dating and measurement.Tree-Ring Bulletni.43:69__75. Liu HB.Shao XM.2003.Reconstruction ofearly—spring temperature of Qinlnig Mountains using tree-ring chronologies.Acta Geographica Sinica,58r6):879__884. Luckman BH.1993.G1acier fluctuation and tree—ring records for the last millennimu in the Canadian Rockies.Quaternary Science Re— views.12:441—45O. Luo L.Wang ZM.Song KS. a1.。2009.Research on the correlation YaJun Wang et a1.,2014/Sciences in Cold andArid Regions,6(1):0030 ̄)036 between NDVI and climatic factors of diferent vegetations in Walpol RE,Myers RH,1985.Probability and Statistics for Engineers northeast China.Acta Botanica Boreall-Occidentalia Sinica,29(4): 8O _808. Sheppard PR,Ⅵ,hite LO.1995.Tree-irng response to the 1978 earth- quake at Stephens Pass,northeastern Califomia.Geology,23: 109—112.D0I: 10.1130/0091-7613(1995)023<0109:TRRTrE> 2_3.CO:2. Stokes MA,Smiley TL,1968.An Introduction to Tree-Ring Dating. Unjversity ofCl1icago Press.Chicago. Swetnam TW Lynch AM,1993.Multicentury,regional-scale patterns of western spruce budworm outbreaks.Ecological Monographs, 63(41:399—424. Tang H Xiao FJ,Zhang Q,et aL,2006.Vegetation change and its response to climate change in the Three-River source region.Ad. vances ni Climate Change Research,2(41:177—1 8O. TreydteKS,SchleserGH,HelleG,eta1.,2006. letwentieth century wasthewettestperiodinnorthemPakistanoverthepastmillennium. Nature,440(27):】l79—1 l82.D0I:1O.1038/nature 04743. Tucker CJ,Fung IV,Keelnig CD,et a1.,1986.Relationship between atmosphericCO2variationsanda satellite—derivedvegetationindex. Nature.319(6050):195-199.DOI:10.1038/319195a0. and Scientists.Macmillna Publishing Company,New York. Wang JS,Y.n QJ,Yang YL,2005.A study of Qinghai grassland productivity using NOAA/AVHRR.Plateau Meteorology,24(1): l17—122. WangYJ,G-aoSY,MaYZ,eta1.,2010.Reconstructingthe annualpre ̄ip— itation variation since 1899 based on tree-ring width in the western Hedong sandy lnad ofNingxi ̄Journal ofArid Land,2f41:286-294. Wang YJ.LuIU.MaYZ, al,2013.Annual vailationinPDSI since 1897 AD ni hte Tengger Desert,Inner Mongolia China,as recorded by rtee.irng data.Journal ofArid Environments.98:2 _26. Wang Z Zhai PM,2003.Clmiate change in drought over northem China during 1950-2000.Acta Geographica Sinica,58(supp1.7、: 61—68. Wigley TML.Brifa KR,Jones PD.1984.On hte average value ofthe correlated time series,wiht applications in dendroclimatology and hydrometeorology.Journal of Climate and Applied Meteorology, 23(2): 201-213.ihtp://dx.doi.org/lO.1175/1520—405o(1984) 023《O201:01:AvOC>2.0.CO;2. Wu XD.1990.Tree Ring and Climate Change.China Meteorological Press,Beijnig.